Highly recommend this as well. Does a great job of helping you build intuition for why things like gradient descent and normalization work. Also gets into the weeds on training dynamics and how to ensure they are behaving properly
> Because the motherfuckers behind public AI interfaces fine-tune them to be as human-like, as rewarding, as dopamine-inducing, as addictive, as possible?
And to think they dont even have ad-driven business models yet
What are some things that made you realize that? Im considering going into research and certainly have some doubts